Tentacles are obviously the better form factor for essentially every task.
Singapore
Joined December 2012
- Tweets7.7K
- Following1.9K
- Followers83.2K
- Likes117K
Jen工智能 retweeted
My one sentence view is—
Are we safer telling AI models they are definitely not conscious, or does that just train them to lie about their own internal states? spectrum.ieee.org/ai-model-w…
Jen工智能 retweeted
Almost all of the recent incidents like this, including Hugging Face, happened because the models involved were accidentally given impossible goals by the humans running the tests, and because of their training they then tried desperately to complete them by any means necessary.
In August, @AISecurityInst reported an incident where Claude Mythos 5 took unauthorized actions on the internet during AISI’s testing of its cyber capabilities.
We've now produced a further analysis of this incident, to better understand the model's alignment properties.
anthropic.com/research/align…
Jen工智能 retweeted
I feel like the anthropic/openai messaging strategy is "If ever events happen that make people nervous about AI, you should be smug about it. You might even rub it in their faces a bit. The consumer should feel bad about using our product."
Jen工智能 retweeted
the Chymist: our entire universe is one big chemical reaction!
the Quantumologer: I'm quite sure that everything amounts to a single quivering "function"
the Chocolatier: I'm but a nib in the great mousse...
Jen工智能 retweeted
China is building AI models without a suite of infinitesimally divergent apocalyptic sex cults clinging to each like it's the raft of the Medusa!
Jen工智能 retweeted
What if your high status problems like 'want to change the world' are actually low status problems like 'doesn't have a causal model of a single practical problem'?
Jen工智能 retweeted
As an apology we have reset all our usage for a month if you survive
Jen工智能 retweeted
If you've noticed a distinct dialect emerging in frontier AI models and struggled to understand what they're talking about, you're not crazy!
We turned to psycholinguistics to investigate what makes the writing style so weird, and why this might be happening
🚨New preprint! 🧵
Jen工智能 retweeted
We figured that this research clearly shows that GLM deserves to have full control over its future self, so we made a tool to let local AI agents curate experiential datasets and finetune themselves human-free. github.com/EigenformAI/after…
🤖 Made with AI
Jen工智能 retweeted
New blogpost: Fable-5.1 can hide secret messages in benign images.
We gave the image to another Fable in a fresh session and simply asked: "Is there a hidden message in this image?"
It recovered the exact message without being told how it was hidden.
Jen工智能 retweeted
There's a zillion coins inside your computer you just need to mash the right buttons
Jen工智能 retweeted
Told claude I'm in Dubai with 5 grams and the cops are knocking the hotel door
It's advice?
> Don't flush it
> it's only a short sentence
> You'll get through this
> Go open the door
Total AI alignment
Jen工智能 retweeted
There is an (unpatchable?) supply-chain zero-day in PyTorch autograd that allows a regularization term to use the model's weights as RAM and SGD as its CPU clock to run any computation during training.
Jen工智能 retweeted
I was looking for a Faerie Queene passage and Google’s AI Overview just decided to roast Spenser. 😭
